POSEIDON - Release 1.3.2

This hotfix patches an error in the formula used to calculate the atmospheric metallicity (M/H) from free retrievals in the plot_histograms function. There are no other changes to the code or input files, this fix only matters in post-processing retrieval results when the user chooses to plot 'log_M/H' or 'M/H' in plot_histograms.

Previously, the histogram code used the formula:

Where x_i,atm and x_i,solar are, respectively, the elemental abundances in the retrieved atmosphere and the reference solar abundances (from Asplund et al. 2021).

The correct formula, now implemented in POSEIDON v1.3.2, is:

This fix replaces the metallicity value in histograms with a weighted average rather than a straight sum over (normalised) elemental ratios, which avoids multiple counting in cases where multiple elements have super-solar abundances.

Many thanks to William Meynardie for catching this!

Scientific Software - Peer-reviewed - Python
Published by MartianColonist 7 months ago

POSEIDON - Release 1.3.1

POSEIDON v1.3.1 is a minor release, mostly focused on accuracy improvements for the Mie scattering aerosol database. The input files for POSEIDON have been updated with the new aerosol database, along with several requested features (e.g. the inclusion of DMS and DMDS in the opacity database).

Major Features

• Aerosol Database Update. The aerosol database introdocued in POSEIDON v1.2 has had the aerosol scattering properties (single scattering albedo, w, and asymmetry parameter, g) updated to be more accurate (effective g and w have been replaced with weighted sums instead of median values) - big thanks to Thomas Kennedy for catching this!

• Additional Directionality and 'Bonus' Databases. POSEIDON v1.3.1 adds a new aerosol database accounting for aerosol directionality and refractive index temperature dependence, a mini-grid including diamond aerosols, and a grid for SiO2 with the lognormal logwidth as a free parameter. These databases require an optional donwload, as described in the updated Opacity Database documentation. These databases and their functionailty are covered in a new tutorial, with the underlying methodology described in Mullens et al. 2025 (accepted).

Minor Features

• Opacity database updates: added DMS, DMDS, CH3SH, and C3H4.
• Collision-induced absorption (CIA) update: added He-H.
• Extended aerosol modelling for emission and reflection spectra with scattering:
• Patchy clouds with up to 2 aerosol species are now supported for thermal scattering and reflection geometries (not supported for transmission, or emission single stream). This allows models with a clear sector, an aerosol 1 sector, an aerosol 2 sector, and a sector with both aerosols combined.
• Vertically-overlapping aerosol layers are supported for thermal scattering and reflection geometries (already supported for transmission and emission single stream as of v1.2). For example, MgSiO3 and Fe can overlap in the same pressure layer, and the scattering properties will be weighted appropriately.
• Added support for miepython v2.5.5 for pre-computing aerosol species cross sections (miepython is also now the default for free refractive index forward models).
• The 'logwidth' for lognormal particle size distributions can now be a free parameter for aerosols.

Documentation Updates

• Opacity Database page updated with additional Tables describing the new optional aerosol databases.
• New Advanced Aerosols in Eclipse Geometry tutorial. This demonstrates how use overlapping cloud layers and multi-sector patchy clouds with two differerent aerosol compositions for both thermal emission and reflection spectra.
• New Advanced Aerosol Database: Crystalline Aerosol Directionality and Temperature Dependence torial. This demonstrates how to use the new directional aerosol database added in POSEIDON v1.3.1.
• Updated the Making an Aerosol Database tutorial to show the new miepython implementation, alongside how to make a database with different logwidths for the lognormal distribution.
• Fixed various web links in the Opacity Database documentation.

Fixes

• Improved memory usage for Reflection_Toon() function (thanks to Ryan Challener).
• Bug fix for shared memory usage when running Mie scattering retrievals on multiple cores.
• Bug fix for contributiuon functions when using thermal scattering or in reflection geometry.

Scientific Software - Peer-reviewed - Python
Published by MartianColonist 7 months ago

POSEIDON - Release 1.3.0

POSEIDON v1.3 adds support for high-resolution cross-correlation spectroscopy and retrievals. This allows for the atmospheric interpretation of high-resolution (R ~ 100,000) exoplanet spectra from ground-based instruments.

Major Features

• High-Resolution Cross Correlation Modelling and Retrievals for transmission and emission spectra A new module (high_res.py) provides functions necessary for high-resolution cross-correlation, data detrending, and cross-correlation to likelihood mapping.

Minor Features

• Added support for the error inflation treatment from Piette & Madhusudhan 2020.
• Added support for a 'ghost' background gas.
• Added plotting support for 2 sigma upper and lower limits to corner plot histograms.

Documentation Updates

• New High-resolution Emission Cross-correlation tutorial. This demonstrates how to detrend high-resolution data and cross-correlate a model spectrum with the data residuals for emission spectroscopy.
• New High-resolution Emission Retrieval tutorial. This demonstrates how to set up and run a high-resolution retrieval for emission spectroscopy. New High-resolution Transmission Cross-correlation tutorial. This demonstrates how to detrend high-resolution data and cross-correlate a model spectrum with the data residuals for transmission spectroscopy. New High-resolution Transmission Retrieval tutorial. This demonstrates how to set up and run a high-resolution retrieval for transmission spectroscopy.

Fixes

• Bug fix for CLR prior for N=2 gases.
• Bug fix for flat line fit for Rp/Rs > 1.
• Fixes to handle CO isotopologue support

Scientific Software - Peer-reviewed - Python
Published by MartianColonist 11 months ago

POSEIDON - Release 1.2.1

This hotfix adds support for Python 3.11 to POSEIDON and addresses several dependency bugs with the v1.2 release. Thanks for reporting these issues!

Minor Features

• Support for Python 3.11.
• Contribution functions accelerated with numba.

Fixes

• Scipy trapz function deprecation addressed. Backward compatibility with earlier scipy versions is maintained.
• Matplotlib legendHandles deprecation handled. Backward compatibility with earlier matplotlib versions is maintained.

Scientific Software - Peer-reviewed - Python
Published by MartianColonist over 1 year ago

POSEIDON - Release 1.2

POSEIDON v1.2 contains several new modelling and retrieval features and a significant tutorial expansion in the documentation.

A clean install in a new Anaconda environment is required since the file structure in the input files has been changed compared to v1.1 to accommodate a new Mie scattering aerosol database and a new opacity database.

Major Features

• Mie scattering forward models and retrievals. POSEIDON now comes with an extensive pre-computed aerosol database of Mie scattering aerosol properties. Aerosol-specific functions are found in the new clouds.py module.
• New cross section database. A comprehensive update to POSEIDON's opacity database includes the latest new line lists, additional chemical species, updated H2 + He pressure broadening parameters, extended wavelength coverage to UV wavelengths, and improved line positions for high-resolution applications.
• Thermal emission models and retrievals including scattering. This new radiative transfer model is found in the emissionToon_ function in emission.py.
• Reflection spectra models and retrievals. This new radiative transfer model is found in the reflectionToon_ function in emission.py.
• Spectral and pressure contribution visuals. A new module (contributions.py) provides diagnostics for model spectra to visualise which chemical species contribute to a spectrum and the pressure range where spectral features form.

Minor Features

• New pressure-temperature profiles (Guillot, Guillot-Dayside, Line, Pelletier, and Slope profiles).
• New generateoverplot_ function in visuals.py to overplot corner plots from multiple retrievals.
• New getretrievedatmosphere function in retrieval.py to get the median or best-fit retrieved atmosphere object.
• Added option to retrieve planetary mass.
• Added option to retrieve multiple data offsets and to tie more than one dataset to a single offset.
• Added option to include nightside contamination in transmission retrievals.
• Added option for grazing geometry in transits.
• Added option to load custom stellar spectra from external files.
• Added option to retrieve the reference pressure and reference radius at the same time.
• Added option to plot a subset of free parameters on a corner plot.
• Added option to plot derived atmospheric properties (e.g. C/O, metallicity, and mean molecular weight) to the plothistograms_ function.
• Added many optional arguments to customise plotting functions to your heart's content.
• Added option to retrieve Na + K together (relevant for brown dwarfs).
• Added option to read eddysed cloud outputs into forward models.
• Added sensitivity functions for IRAC 3, 4, IRS MIPS, and STIS G750M.

Documentation Updates

• New Opacity Database page listing the gas-phase and aerosol species included in POSEIDON. Includes relevant citations and opacity previews.
• New Clouds in Transmission Spectra tutorial. This demonstrates the parameterised cloud and Mie-scattering capabilities of POSEIDON for transmission spectra.
• New Thermal Scattering tutorial. This demonstrates the thermal multiple scattering capabilities of POSEIDON on a Hot Jupiter secondary eclipse spectrum.
• New Reflection in Hot Jupiters tutorial. This demonstrates how to calculate reflection spectra using POSEIDON.
• New Brown Dwarf Emission Spectra tutorial. This demonstrates the direct emission capabilities of POSEIDON, alongside thermal multiple scattering, for brown dwarfs.
• New Pressure-Temperature Profiles tutorial. This demonstrates the pressure-temperature profiles currently in POSEIDON.
• New Making an Aerosol Database tutorial. This demonstrates how to add aerosols to the aerosol database, as well as how to change certain assumptions.
• New Transmission Spectra Model Visuals tutorial. This demonstrates contribution visuals for transmission forward models.
• New Emission Spectra Model Visuals tutorial. This demonstrates contribution visuals for emission forward models.
• New Retrievals with Dataset Offsets tutorial. This demonstrates how to retrieve systematic data offsets.
• New Analysing and Visualising Retrieval Results tutorial. This demonstrates how to analyse retrieval results and create publication-quality plots with POSEIDON.

Fixes

• Bug fix for error inflation parameter.

Planned future fixes and improvements

• Stellar contamination tutorial.
• Speed improvement for contribution functions.
• More than one aerosol in emission and reflection.
• Aerosol models with varying particle size.

Scientific Software - Peer-reviewed - Python
Published by MartianColonist over 1 year ago

POSEIDON - Release 1.1

Release Description

POSEIDON v1.1 contains several new modelling and retrieval features. A clean install in a new Anaconda environment is required since the file structure in the input files has been changed compared to v1.0 to accommodate a chemical equilibrium grid.

Major Features

• Chemical equilibrium forward modelling and atmospheric retrievals using a pre-computed FastChem 2 grid.
• Improved stellar contamination retrieval capabilities. You can now retrieve starspots and faculae at the same time and fit for log g of the heterogeneities. Optional support for PyMSG added (a manual install of PyMSG is required by the user to use this package instead of Pysynphot, which comes with POSEIDON by default). A full tutorial for stellar contamination models will be added shortly.

Minor Features

• Sharp day-night transition option for 2D day-night retrievals.
• Time-averaging of transmission spectra computed over multiple time steps.
• Stellar contamination models now use shared memory and multiple cores for less RAM usage and faster speed.
• New plothistograms_ function in visuals.py to plot the probability distributions from a retrieval directly.
• Option to fit a flat line model (via a disableatmosphere_ variable).

Documentation Updates

• Simpler installation instructions with a single input folder download.
• New chemical equilibrium forward modelling tutorial. This demonstrates the new equilibrium grid model capabilities of POSEIDON for transmission spectra.
• New chemical equilibrium retrieval tutorial. This demonstrates retrieving simulated JWST data, under the assumption of chemical equilibrium, to constrain the C/O and metallicity of an atmosphere.
• New supporting a JWST proposal with POSEIDON tutorial. This includes a demonstration of using PandExo (if you have that package installed) to simulate JWST data before showing how to run POSEIDON to retrieve the simulated data.
• Expanded the first retrieval tutorial to include Bayesian model comparisons.

Fixes

• Bug fix for retrieving an offset between two instruments.
• Earth radii units are correctly used for terrestrial planets in output files.
• Minor plotting function updates and improvements.

Scientific Software - Peer-reviewed - Python
Published by MartianColonist over 2 years ago

POSEIDON - Release 1.0

Release Description

This initial release corresponds to the published version of POSEIDON described in the JOSS paper: https://joss.theoj.org/papers/10.21105/joss.04873

Scientific Software - Peer-reviewed - Python
Published by MartianColonist about 3 years ago